1
Improving Sepsis Care with Data Analytics
Session 243, February 14, 2019
Mischa Adams, MSN, RN, CCRN, Clinical Standards Coordinator
Sarah Jenson, Analytics Director
2
Mischa Adams MSN, RN, CCRN Clinical Standards Coordinator
Consulting Fees: Cheetah Medical
Sarah Jenson, MS
Has no real or apparent conflicts of interest to report.
Conflict of Interest
3
• History of Sepsis at Allina
• Program Development
• Data Access
• Improvement
• Future
Agenda
4
• LO1: Describe how a data-driven approach to sepsis care can
improve patient outcomes
• LO2: Describe how a healthcare organization can use data to
monitor performance and gain clinician buy-in
• LO3: Explain why an enterprise data warehouse with dedicated
septic analytics is essential to increase early identification of
sepsis and adherence to treatment bundles
Learning Objectives
5
Allina Health is the Minnesota Region’s
Broadest Care System
About Allina:
• 85 clinic sites and ambulatory care
centers.
• 5,000 physicians.
• 3.0 million+ clinic visits.
• 12 hospitals.
• 103,000+ inpatient hospital
admissions.
• $4B in revenue.
• 31% Twin Cities inpatient market
share.
6
In the U.S., over 1.5 million people are treated for sepsis annually.
One in four people with sepsis die, making improving early
identification and providing patients with timely treatment a top
priority.
Sepsis
Seymour CW, Rea TD, Kahn JM, Walkey AJ, Yealy DM, Angus DC: Severe sepsis in pre-hospital emergency care: analysis of incidence, care, and outcome. Am J Respir Crit Care Med 2012, 186(12):1264-1271.
8
• Clinician awareness of assessment and interventions required for
sepsis core measures
• Multiple, often conflicting, procedures to activate a sepsis alert (an
early warning notification within the EMR indicating the patient
meets criteria for sepsis)
• Fourteen different sepsis order sets
• Understanding documentation processes to accurately reflect
patient care
• Lack of data visibility to understand performance
Challenges
9
Sepsis at Allina Health
0%
5%
10%
15%
20%
25%
30%
0
50
100
150
200
250
300
350
Mortality Rate
Sepsis Visit Volume
Sepsis Mortality over Time
Visits In-Hospital Mortality Rate
AATP
Nurse
Coordinators
Hired
Ordersets
Updated
Fluid End
Time
Mandatory
Nursing
Education
Released
Inpatient
Electronic
Screening
10
• Develop Sepsis Program
– Leaders from Emergency Medicine, Intensivists,
Hospitalists, and Nursing
– Ancillary Pharmacy and Laboratory teams included as
necessary
• Sepsis Program Coordinators are registered nurses (RN) with
both clinical expertise and performance improvement skills.
– Monitor performance and facilitate process improvement
efforts at the point of care
– Ensuring protocols are in alignment with current best
practices.
– Develop and Deliver Education
– Deliver Provider feedback
Support
12
• More than 85 percent of Allina Health Sepsis patients had
infection with organ dysfunction identified in the ED
– Physician education timely identification and treatment
– Nursing education focused on the early identification of
sepsis at triage and rapid initiation of sepsis orders.
Focus
0
200
400
600
800
1,000
Sepsis Volume
ED Identification IP Identification
13
• Improving identification of sepsis patients for coding purposes and
provider education
• Evaluation of Best Practice Alerts (BPAs)
• Automated identification of organ dysfunction
• Identifies gaps and variation in sepsis care
• Assess Protocol-based interventions
• Provide feedback to physicians about how adherence to bundles
improved outcomes
Data Unlocks Improvements
14
Please use this blank slide if more space is
required for charts, graphs, etc.
Remember to delete this slide, if not needed.
16
• Early identification has reduced the number of patients developing
Septic Shock
Reducing Septic Shock
0
5
10
15
20
25
30
35
Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4
2014 2015 2016 2017 2018
Rate per 100 Sepsis Admissions
Sepsis Volume -
Rate per 100 Sepsis Admissions
Septic Shock Severe Sepsis
17
• Increasing control and decreasing overall average time from
identification of sepsis to antibiotic order.
Antibiotic Timing
0
100
200
300
400
500
600
700
Avg Time Zero to Abx Order UCL LCL Mean
18
• Demonstrated improvement in Fluid Administration timing
Fluid Administration
0
20
40
60
80
100
120
140
Avg Time Zero to Fluid Complete UCL LCL Mean
19
20
21
• 18.6% Relative reduction in mortality rate and 10.9% relative
reduction in hospital Length of Stay (LOS) for all patients with
sepsis
• 30.3% relative reduction in morality rate and 18.4% relative
reduction in hospital LOS for patients with sever sepsis and septic
shock
• $1.1 million in annual cost savings, the result of efficiencies and
substantial reductions in hospital LOS for patients with severe
sepsis or septic shock
Results
22
• 2019 shift focus to Inpatient presentation of Sepsis
• Ongoing monitoring of ED Shock cases
Future
23
• Please complete online session evaluation
Questions
Sarah.Jenson@healthcatalyst.com
Sarah Jenson
Mischa Adams
Mischa.Adams@allina.com
